Build an AI Appointment Caller with OpenAI Realtime API & Google Calendar

[@Hhhcghvcg]

Very good video 👌

[@BartSlodyczka]

Thanks legend!

[@redwanshaikh9117]

Thanks dude, you are making it great and simple.

[@BartSlodyczka]

Thank you legend 🤝

[@РусланНагимов-д7д]

another amazing video

[@BartSlodyczka]

Thank you :)

[@JonathanNatakusuma-n5c]

Thanks Bart much appreciated!

[@BartSlodyczka]

thank you 💪

[@thenoblerot]

Great channel, Bart - commenting for the algorithm!

[@BartSlodyczka]

Thank you legend 🏆

[@2626tags]

Great, this is very helpful.

[@BartSlodyczka]

Thanks!

[@p4r7h-v]

banger video

[@BartSlodyczka]

my man! 💪

[@JanMdwoe]

Great video, what can I use as an alternative to Replit? Im not allowed to use the plattform because of EU Data Privacy. What do I need instead - could it be done with DigitalOcean?

[@BartSlodyczka]

Thanks! I think so, although I haven't used any other platforms aside from Replit as Replit is super easy for me to use. But I think AWS or Google Cloud would also work :)

[@zehaankhan1]

ok great... So how many calls can it place in at a time.... concurrency?... or is it one call at a time?

[@BartSlodyczka]

Not sure how many, I think it depends on how you deploy, so if the current deployment settings limit calls, you could upgrade the server to handle more volume. But in truth I haven't tested max calls on this exact set up

[@kamalkamals]

there is no alternative of twilio ?

[@BartSlodyczka]

Haven't looked into it much so not sure 🙏

[@derherrdirector]

there are plenty alternatives to twilio. Just search for SIP services in your country.

[@itayhilel2168]

Can you help us with an alternative to twilio?

[@BartSlodyczka]

Give me some suggestions and I'll look into it :)

[@meanfitness]

Can the bot actually take an order for say a fast food restaurant and send to owner? Also send payment link to customer?

[@BartSlodyczka]

Yeah for sure, as long as the restaurants booking tool supports api connectivity, or ideally it has modules in make.com. Would just need to build out the appropriate function calls

[@AbeelIshrat]

You said for the transcript_webhook_url and booking_webhook_url, we have to copy the URLs from the First Webhook in the diagram and paste them, but you have 2 different URLs in your code Can you guide furthur? That would be very helpful Thankyou

[@BartSlodyczka]

the transcript webhook is used at the end of the conversation and is where we send the transcript of the conversation. In my Part 3 video I give you a make.com scenario that can be used to help process the transcript. And yeah this is a seperate webhook to the booking webhook. For the booking webhook, use the make.com scenario that i supply in this video (can find in the gumroad). To summarise, use two different webhooks! Hope this helps :)

[@AbeelIshrat]

@@BartSlodyczka Thankyou soo much you are awsomeee

[@cb4623]

Ya buddy!

[@BartSlodyczka]

Let's gooo!!

[@danielpistola]

I wanted to ask.... Is the realtime API more reliable now?

[@BartSlodyczka]

I think it is better yes, but there are still some niggly issues that pop up sometimes. Like sometimes the ends of words are cut off (when the AI agent is responding). I've read about this as well in the openAI forums so I think it is on the radar and will be taken care of soon :)

[@shib_ai]

👍

[@BartSlodyczka]

💪💪